Job Role: Machine Learning Engineer

A machine learning engineer (MLE) is a key part of a team working for a business that wants to benefit from machine learning and artificial intelligence functionality.

In general, these professionals are the guardians of some of the most powerful technologies around.

Let’s look at more on what MLEs do from machine learning professionals in the tech field who spoke to us about what it means to fill one of these job roles.

Some of the Machine Learning Engineer Basics

At a very basic level, a machine learning engineers has to understand the nuts and bolts of how these projects get put together, and how they get shepherded toward completion. (Read Machine Learning 101.)

“A machine learning engineer is responsible for getting the machine learning based solution up and running,” says Rosaria Silipo, Ph.D., principal data scientist at KNIME. “This involves some programming to integrate the machine learning algorithm into the project and some IT experience to deploy the final solution into production.”

Silipo suggests that practical real-world experience in machine learning will go a long way, since it will address the ability to troubleshoot.

“A machine learning engineer is expected to have enough machine learning knowledge to understand the role of the algorithm in the application and solve potential pitfalls and malfunctioning,” Silipo says.

That's where some of these programming and mathematical skills come in. MLEs often want to know programming languages like Python, and are pretty conversant with back-of-the-napkin math tricks. (Read The Debate Between R and Python.)

More Input from Machine Learning Engineers

Some of those who have a first row seat to the ways that companies work with machine learning engineers also echo the sentiment that these professionals are uniquely powerful in the business world.

“Machine learning engineers are top-echelon programmers who design self-running software that can learn and apply knowledge on autopilot without humans chiming in,” says Maciej Baranowski, Machine Learning andMarketing Automation Specialist at Zety.

Baranowski suggests machine learning engineers are very much in demand based on specific company goals that have to do with time efficiencies.

“Companies have a haystack of data sets to process and if they do things manually, it’ll be like attempting to watch 13 Star Trek movies in one hour,” Baranowski says.

Alex Bekker, Head of Data Analytics Department at ScienceSoft, has a more prosaic description of what machine learning engineers do.

“Machine learning engineers are responsible for preparing the data required to train an ML model (for example, they augment data and reduce noise) and working with ML models,” Bekker says, describing how things work in his own team. “The latter implies designing the models, choosing relevant activation and optimization functions, tuning the models' hyperparameters, training and retraining the models, making sure that the model differentiates signals from noise, and more. “

Working with machine learning models, he says, involves a variety of high-level tasks, again, covering the lifecycle of development.

Understanding Some of the Risks

Here's another way to understand what the machine learning engineer does, and to realize their importance in the enterprise production of machine learning projects.

Recently, the U.S. Department of Justice bringing charges against former Google machine learning engineer Anthony Levandowski claiming that he took intellectual property in the form of trade secrets with him to Uber.

Here's an example where a self-driving car engineer had so much power over the process that he was able to effectively steal vital trade secrets from an employer.

This illustrates how companies have to essentially trust machine learning engineers and give them certain amounts of power over the process and access to sensitive information. They are so central to the process that firms have to look carefully at vetting them and putting requisite controls in place.

The Day-to-Day Life of a Machine Learning Engineer

David Khachaturov, a machine learning engineer at speech recognition technology company Speechmatics, gave us a more detailed picture of how machine learning engineers might work in practice.

“A large part of being a machine learning engineer involves staying on top of the latest developments in the AI/ML field – whether that’s scouring the GitHub Trending Page, frantically refreshing arXiv or spending too much time scrolling through reddit.com/r/MachineLearning,” Khachaturov says. “They attempt to repurpose research-oriented technologies for more commercial contexts. This repurposing commonly involves adding the extra functionality and usability that the original academic work will often lack.”

Sometimes, he says, big breakthroughs in machine learning happen quickly.

“At these points, machine learning engineers often kick into high gear and quickly try to develop a proof-of-concept (PoC) that can add extreme value to their business and customers,” he adds. “If the PoC works as intended, it can be fleshed out into a whole new product. This can be very exciting in terms of opening up new opportunities both internally for the company and for the end customers due to the increased technical leverage gained by using cutting-edge machine learning.”

Khachaturov also goes back to what many others have also told us about ML engineers, that some of the more tedious work involves cleaning and managing data. There are many rote tasks related to this ongoing need for data curation and handling, and even though some of it can be automated, the rest of it isn’t going away any time soon.

“A machine learning engineer should be ready for anything,” Khachaturov says, “From trying to get data into the right input format to writing custom web scrappers. “

What We've Learned

Hopefully this gives you a bit of a better picture of how ML engineers spend their time. This important job role is one that will continue to get a lot of attention as we plumb the capabilities of ML to transform our business and our lives.